CN113220801B

CN113220801B - Structured data classification method, device, equipment and medium

Info

Publication number: CN113220801B
Application number: CN202110538741.5A
Authority: CN
Inventors: 刘焱; 姚兴
Original assignee: Alipay Hangzhou Information Technology Co Ltd
Current assignee: Alipay Hangzhou Information Technology Co Ltd
Priority date: 2021-05-17
Filing date: 2021-05-17
Publication date: 2022-07-29
Anticipated expiration: 2041-05-17
Also published as: CN113220801A

Abstract

The embodiment of the specification provides a structured data classification method, a structured data classification device, equipment and a medium. The method comprises the following steps: performing sampling operation on one or more tables stored in a structured database so as to obtain a field set corresponding to the tables; obtaining field information corresponding to each field in the field set, and performing scoring operation on the field information by using a pre-configured scoring rule or a scoring model to obtain a scoring result corresponding to the field; wherein the scoring result can obscure categories characterizing the field; training a preset deep learning model according to the scoring result corresponding to the field to obtain a trained deep learning model; and predicting the field of the structured data of the unknown class by using the trained deep learning model, and determining the class corresponding to the structured data according to the prediction result.

Description

Structured data classification method, device, equipment and medium

Technical Field

The present disclosure relates to the field of data processing technologies, and in particular, to a structured data classification method, apparatus, device, and medium.

Background

With the wide application of the internet and big data technology, the network data security is highly valued by all levels of governments and enterprises and public institutions, and especially with the increasing importance of the country on the user privacy protection, various laws and regulations for data security and privacy protection are introduced. The internet company collects various user data in the normal business process, for example, personal privacy data such as a mobile phone number, an identification number, a bank card number and the like under the condition of obtaining user authorization. When the security privacy protection is performed on the user data, the user data stored by enterprises needs to be classified and classified, and then the data with different security levels is further protected according to the national regulations, so that the visible data classification is the basis for the user privacy protection.

In the prior art, the current algorithm for classifying structured data is mainly based on expert rules of relevant information in a database and a manual marking mode for judgment, and the expert rules are difficult to contain all possible situations, difficult to balance on accuracy and recall rate and extremely high in maintenance cost; the manual marking mode has the problems of large workload, inextensibility and the like.

Based on the prior art, a classification scheme with higher accuracy and recall rate, no need of a large amount of manual marks and expandability needs to be provided.

Disclosure of Invention

Embodiments of the present description provide a method, an apparatus, a device, and a medium for structured data classification, so as to solve the problems of low accuracy and recall rate, high maintenance cost, large number of manual marks required, and no expandability in the prior art.

In order to solve the above technical problem, the embodiments of the present specification are implemented as follows:

an embodiment of the present specification provides a structured data classification method, which includes:

performing sampling operation on one or more tables stored in a structured database so as to obtain a field set corresponding to the tables;

obtaining field information corresponding to each field in the field set, and performing scoring operation on the field information by using a pre-configured scoring rule or a scoring model to obtain a scoring result corresponding to the field; wherein the scoring result can obscure categories characterizing the field;

training a preset deep learning model according to the scoring result corresponding to the field to obtain a trained deep learning model;

And predicting the field of the structured data of the unknown class by using the trained deep learning model, and determining the class corresponding to the structured data according to a prediction result.

An embodiment of the present specification provides a structured data classification apparatus, where the apparatus includes:

the sampling module is used for performing sampling operation on one or more tables stored in a structured database so as to obtain a field set corresponding to the tables;

the scoring module is used for acquiring field information corresponding to each field in the field set, and performing scoring operation on the field information by utilizing a pre-configured scoring rule or a scoring model to obtain a scoring result corresponding to the field; wherein the scoring result can obscure categories characterizing the field;

the training module is used for training a preset deep learning model according to the scoring result corresponding to the field to obtain a trained deep learning model;

and the prediction module is used for predicting the field of the structured data of the unknown class by using the trained deep learning model and determining the class corresponding to the structured data according to the prediction result.

An embodiment of the present specification provides a structured data classification device, including:

At least one processor; and the number of the first and second groups,

a memory communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory stores instructions executable by the at least one processor to enable the at least one processor to:

and predicting the field of the structured data of the unknown class by using the trained deep learning model, and determining the class corresponding to the structured data according to the prediction result.

Embodiments of the present specification provide a computer readable medium having computer readable instructions stored thereon, which are executable by a processor to implement a structured data classification method as described above.

The embodiment of the specification adopts at least one technical scheme which can achieve the following beneficial effects:

sampling one or more tables stored in a structured database so as to obtain a field set corresponding to the tables; obtaining field information corresponding to each field in the field set, and performing scoring operation on the field information by using a pre-configured scoring rule or a scoring model to obtain a scoring result corresponding to the field; the scoring result can fuzzily represent the category of the field; training a preset deep learning model according to the scoring result corresponding to the field to obtain a trained deep learning model; and predicting the field of the structured data of the unknown class by using the trained deep learning model, and determining the class corresponding to the structured data according to the prediction result. Based on the scheme, the scoring rule or the scoring model is utilized to score the field for training the deep learning model, the obtained scoring result can be used for fuzzily representing the category of the field, namely, noise possibly exists in the scoring result, and then the deep learning model is trained through the scoring result containing the noise, so that how to filter the noise is automatically learned, and correct selection is made, so that when the trained deep learning model is utilized to identify the field of unknown category, better accuracy and recall rate can be obtained, and the method does not need a large amount of manual marks and has expandability.

Drawings

In order to more clearly illustrate the embodiments of the present specification or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, it is obvious that the drawings in the following description are only some embodiments described in the present specification, and for those skilled in the art, other drawings can be obtained according to the drawings without any creative effort.

FIG. 1 is a flow chart illustrating a method for structured data classification according to an embodiment of the present disclosure;

FIG. 2 is a schematic structural diagram of a structured data classification apparatus corresponding to FIG. 1 provided in an embodiment of the present disclosure;

fig. 3 is a schematic structural diagram of a structured data classification apparatus corresponding to fig. 1 provided in an embodiment of the present specification.

Detailed Description

In order to make those skilled in the art better understand the technical solutions in the present specification, the technical solutions in the embodiments of the present specification will be clearly and completely described below with reference to the drawings in the embodiments of the present specification, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any inventive step based on the embodiments of the present disclosure, shall fall within the scope of protection of the present application.

As mentioned above, internet companies collect various user data during normal business processes, and the user data is usually stored in a relational database (i.e., a structured database) in a table form, and is also called structured data. When privacy protection is performed on user data, the user data with different security levels need to be protected according to national regulations, and the user data with different types often correspond to different levels, so that classification of the user data is a basic task for privacy protection of users.

The current main structured data classification algorithm is mainly divided into two types, the first type is that a regular expression based on field characteristics and field annotation information in a database is classified, no matter the field characteristics or the field annotation information, the regular expression belongs to an expert rule, all possible situations are difficult to be exhausted by utilizing the regular expression, the balance between accuracy and recall rate is difficult, and the maintenance cost is extremely high; secondly, classification is performed in a manual marking mode, which is heavy in workload and not extensible, and particularly in internet companies, the manual marking mode can hardly be used due to rapid business development and frequent changes of database structures.

It should be noted that, in the following embodiments of this specification, data security and privacy protection are described as application scenarios, and data related to user privacy (such as user name, mobile phone number, identification card number, bank card number, and the like) is mainly stored in the structured database, but in practical application, the scheme is not limited to the technical scenario of data security and privacy protection, and is also applicable to the scheme when data classification is performed on non-user privacy data; in addition, the execution main body of the scheme may be a server, or may be other terminal devices, such as a PC, a tablet computer, a mobile terminal, and the like, and the relevant operations may be implemented by codes or programs running in the server or the terminal devices. The following describes a scheme of the present specification in detail, taking a classification scenario of user privacy data as an example.

Fig. 1 is a schematic flowchart of a structured data classification method provided in an embodiment of the present specification, where the method specifically includes the following steps:

in step S110, a sampling operation is performed on one or more tables stored in the structured database, so as to obtain a field set corresponding to the table.

In one or more embodiments of the present description, the structured database may be a database adopting a distributed structure, for example, a distributed database adopting a block chain technology, and when the user privacy data is stored by adopting a block chain, the user privacy data may be stored in a distributed node of the block chain in a form of a table. In practical applications, the structured database is not limited to the database with a distributed structure, and an independent database or a clustered database is also suitable for the scheme.

Further, in this embodiment of the present specification, the data stored in the structured database may be personal privacy data of the user collected by the enterprise in the normal business operation process, and therefore, when collecting this part of data, an authorization request needs to be sent to the user first, and after obtaining the authorization of the user, the user data is obtained.

In addition, when sampling the tables in the structured database, it is also necessary to send a data sampling request to a collector of user data (such as an internet enterprise), and when obtaining the authorization of the internet enterprise, the authorized tables are sampled from the structured database, that is, the authorization status of the tables is determined, and the sampling operation is performed on the authorized tables.

When the table is sampled, the granularity of the sampled data may reach a field level, a field may be understood as a column in the table, and fields under the same column belong to a type of field, for example, a user identifier, a user name, a user identification number, a user card number, and the like in the user privacy data table all correspond to a type of field or are referred to as a field. For a field of the same class in the table, at least a preset number of rows need to be sampled (for example, K rows need to be sampled), and assuming that X types of fields in the table need to be sampled, that is, X types of fields need to be sampled, the number of fields in a field set formed by all the collected fields is K X, that is, sampling will generate K X data sets, and each data set corresponds to one field value in the table.

In a specific embodiment of the present specification, the field information as shown in table 1 below (web _ user table);

table 1web _ user table

userId	name	card number
			1	zhangsan	11022819901026xxxx
2	xiaoming	13018219890120xxxx
			3	laowang	33080319890917xxxx
……	……	……

The userId, the name and the card number in the web _ user table belong to three different fields respectively, when the table is sampled, K (K is more than or equal to 1000) row fields respectively corresponding to the three fields can be collected, and when the value of K is 1000, the number of data sets corresponding to the collected fields is K X1000X 3.

In step S120, field information corresponding to each field in the field set is obtained, and a pre-configured scoring rule or a scoring model is used to perform a scoring operation on the field information to obtain a scoring result corresponding to the field; wherein the scoring result can obscure the category characterizing the field.

In one or more embodiments of the present specification, before scoring each field, field information corresponding to each field in the field set obtained in step S110 needs to be obtained first, where the field information includes a value of the field and comment information of the field; the value of the field refers to the feature described by the field, for example, zhangsan in table 1 above is the field value under the name field, and the comment information of the field refers to the comment description made for each type of field in the table.

Specifically, in the embodiment of the present specification, a value of a field corresponding to each field and annotation information of the field are used as input, a preset scoring rule or a scoring model is used to score the input, scoring results corresponding to each field are obtained, and the scoring results are uploaded to nodes in a block chain; the contents and principles of the scoring rules and the scoring models are described in detail below with reference to embodiments, which specifically include the following:

the preset scoring rule or the pre-configured scoring model is a scoring rule or a scoring model configured according to the category of the known field; wherein, the first and the second end of the pipe are connected with each other,

when the field types are known to be multiple, generating a rule set according to the regular expressions configured for each field type, and scoring each field by using the rule set to obtain a scoring result consisting of multiple dimensions;

the scoring model comprises a machine learning classification model or a natural language processing model, and is used for scoring and predicting field information corresponding to the field to obtain probability values of the field belonging to a field category.

In this embodiment of the present specification, when a field is marked in a regular expression manner, one regular expression may be configured for each field type according to a known field type, and the regular expressions corresponding to all the field types are placed in one rule set. For example, common field types such as a mobile phone number, an identification number, a bank card number, etc. may be covered, and the field type in the rule set is set to N. When the rule set is used for scoring and judging the sampled fields, a scoring result with a dimension of N can be obtained for each field, and then the data set of the scoring result output by the fields in the field set is [ K X, N ].

Taking a regular expression matching the mobile phone number as an example, ^ ((13[0-9]) | (14[5,7]) | (15[0-3,5-9]) | (17[0,3,5-8]) | (18[0-9]) |166|198|199| (147)) \\ \ d {8}, based on the regular expression, a scoring result corresponding to the field can be calculated, and if a scoring rule is hit, the scoring result is 1, otherwise the scoring result is 0. It should be noted that, in the embodiments of the present specification, a regular expression may also be set for the comment information of the field, and the regular expression corresponding to the field and the regular expression of the comment information of the field are placed in the same rule set.

When the scoring judgment is carried out in a scoring model mode, taking a natural language processing model as an example, the values of the fields and the annotation information of the fields can be used as parameters to be input into the natural language processing model, and the natural language processing model calculates a scoring function according to the parameters to obtain a scoring result; for example, when determining whether a field is an address, a probability value of the field belonging to the address, i.e., a probability value between 0 and 1, may be returned.

In practical application, besides the regular expression and machine learning classification model, the method can also adopt the keyword list and other methods, and the field scoring method does not form the limitation of the scheme.

In step S130, a predetermined deep learning model is trained according to the scoring result corresponding to the field, so as to obtain a trained deep learning model.

In one or more embodiments of the present specification, the predetermined deep learning model is a pre-constructed codec, and the codec is a custom model based on a deep learning framework; wherein the content of the first and second substances,

the structure of the codec comprises three hidden layers, the number of neurons between a first hidden layer and a third hidden layer is the same, and the number of neurons of the second hidden layer is smaller than that of the neurons of the first hidden layer and the third hidden layer; and adding white Gaussian noise between the first hidden layer and the second hidden layer.

In a specific embodiment of the present specification, the codec may be a model defined by a deep learning network model framework, and the codec adopts a structure of three hidden layers, where the number of neurons in the first hidden layer and the third hidden layer is X, the number of neurons in the second hidden layer is Y, and Y should be smaller than X. In practical application, in order to further improve the denoising capability, a white gaussian noise with an average value of 0 and a standard deviation of s (0 < s < 0.2) can be randomly added between the first hidden layer and the second hidden layer.

In the custom codec of the embodiment of the present specification, the first hidden layer to the second hidden layer may be considered as an encoding process for abstracting and extracting high-level features, and the second hidden layer to the third hidden layer is a decoding process for feature restoration. The gaussian white noise is added in such a way that a function value is randomly added to the data input between the first hidden layer and the second hidden layer, which function value can act as a gaussian noise generator.

Further, before the codec is trained, randomly dividing the data set [ K X, N ] obtained after the steps S110-S120 are performed into a training set and a test set, training the constructed custom codec by using the training set, and setting a loss function as a Mean Square Error (MSE) between an input of the first hidden layer and an output of the third hidden layer in the training process of the codec.

In the deep learning model training process, a loss function is defined, for example, a mean square error between an input of a first hidden layer and an output of a third hidden layer is used as the loss function, so that the input of the first hidden layer and the output of the third hidden layer are ensured to be as close as possible, namely, the loss function is made to be smaller and smaller in the deep learning model training process, the loss function is reversely transferred to adjust parameters of the model in each training process, and the whole model training process can be regarded as a process of making the loss function smaller and smaller.

It should be noted that the codec trained in the above manner can automatically filter out dirty data (e.g. fields that do not belong to a known type, etc.) and noise, and avoid the influence of noise on model prediction. According to the scheme, the field is scored for the first time by using a scoring function (namely, a scoring rule or a scoring model), and the scoring result can only fuzzily represent the category to which the field belongs but not an accurate classification result because the regular expression may have the condition of false report or missed report. Considering the situation, the false alarm or the false negative existing in the first scoring result is regarded as a kind of noise, and a Gaussian noise generator is additionally added to train the deep learning self-codec, so that how to filter the noise is automatically learned, and a correct choice is made. Because the scheme can tolerate the false report or the missing report of different regular expressions, better accuracy and recall rate can be obtained when the structured data is classified based on the scheme, and the scheme does not need to rely on a large amount of manual marks and has expandability.

Based on the same idea, an embodiment of the present specification further provides a structured data classification apparatus, and as shown in fig. 2, a schematic structural diagram of the structured data classification apparatus provided in the embodiment of the present specification is provided, the apparatus 200 mainly includes:

A sampling module 201, configured to perform sampling operation on one or more tables stored in a structured database, so as to obtain a field set corresponding to the table;

a scoring module 202, configured to obtain field information corresponding to each field in the field set, and perform a scoring operation on the field information by using a pre-configured scoring rule or a scoring model to obtain a scoring result corresponding to the field; wherein the scoring result can obscure categories characterizing the field;

the training module 203 is used for training a preset deep learning model according to the scoring result corresponding to the field to obtain a trained deep learning model;

and the prediction module 204 is configured to predict a field of the structured data of an unknown class by using the trained deep learning model, and determine a class corresponding to the structured data according to a prediction result.

Further, the field information includes a value of the field and annotation information of the field, and the scoring module 202 is further configured to:

and taking the value of the field corresponding to each field and the comment information of the field as input, scoring the input by utilizing a pre-configured scoring rule or a pre-configured scoring model to obtain a scoring result corresponding to each field, and uploading the scoring result to a node in the block chain.

Further, the pre-configured scoring rule or scoring model is configured according to the category of the known field; wherein the content of the first and second substances,

Further, the preset deep learning model is a pre-constructed codec, and the codec is a custom model based on a deep learning framework; wherein the content of the first and second substances,

Further, the training module 203 is further configured to:

in the training process of the codec, a loss function is set as a mean square error between an input of the first hidden layer and an output of the third hidden layer.

Based on the same idea, an embodiment of the present specification further provides a device corresponding to the foregoing method, and fig. 3 is a schematic structural diagram of a structured data classification device corresponding to fig. 1 provided in the embodiment of the present specification. As shown in fig. 3, the apparatus 300 may include:

at least one processor 310; and the number of the first and second groups,

a memory 330 communicatively coupled to the at least one processor; wherein the content of the first and second substances,

the memory 330 stores instructions 320 executable by the at least one processor 310 to enable the at least one processor 310 to:

Based on the same idea, embodiments of the present specification further provide a computer-readable medium corresponding to the above method, where the computer-readable medium has stored thereon computer-readable instructions, and the computer-readable instructions can be executed by a processor to implement the above structured data classification method.

The foregoing description has been directed to specific embodiments of this disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous.

The embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the embodiments of the apparatus, the electronic device, and the nonvolatile computer storage medium, since they are substantially similar to the embodiments of the method, the description is simple, and the relevant points can be referred to the partial description of the embodiments of the method.

The apparatus, the electronic device, the nonvolatile computer storage medium and the method provided in the embodiments of the present description correspond to each other, and therefore, the apparatus, the electronic device, and the nonvolatile computer storage medium also have similar advantageous technical effects to the corresponding method.

In the 90 s of the 20 th century, improvements in a technology could clearly distinguish between improvements in hardware (e.g., improvements in circuit structures such as diodes, transistors, switches, etc.) and improvements in software (improvements in process flow). However, as technology advances, many of today's process flow improvements have been seen as direct improvements in hardware circuit architecture. Designers almost always obtain the corresponding hardware circuit structure by programming an improved method flow into the hardware circuit. Thus, it cannot be said that an improvement in the process flow cannot be realized by hardware physical modules. For example, a Programmable Logic Device (PLD), such as a Field Programmable Gate Array (FPGA), is an integrated circuit whose Logic functions are determined by programming the Device by a user. A digital system is "integrated" on a PLD by the designer's own programming without requiring the chip manufacturer to design and fabricate application-specific integrated circuit chips. Furthermore, nowadays, instead of manually making an Integrated Circuit chip, such Programming is often implemented by "logic compiler" software, which is similar to a software compiler used in program development and writing, but the original code before compiling is also written by a specific Programming Language, which is called Hardware Description Language (HDL), and HDL is not only one but many, such as abel (advanced Boolean Expression Language), ahdl (alternate Hardware Description Language), traffic, pl (core universal Programming Language), HDCal (jhdware Description Language), lang, Lola, HDL, laspam, hardward Description Language (vhr Description Language), vhal (Hardware Description Language), and vhigh-Language, which are currently used in most common. It will also be apparent to those skilled in the art that hardware circuitry that implements the logical method flows can be readily obtained by merely slightly programming the method flows into an integrated circuit using the hardware description languages described above.

The controller may be implemented in any suitable manner, for example, the controller may take the form of, for example, a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro) processor, logic gates, switches, an Application Specific Integrated Circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include, but are not limited to, the following microcontrollers: ARC625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, the memory controller may also be implemented as part of the control logic for the memory. Those skilled in the art will also appreciate that, in addition to implementing the controller as pure computer readable program code, the same functionality can be implemented by logically programming method steps such that the controller is in the form of logic gates, switches, application specific integrated circuits, programmable logic controllers, embedded microcontrollers and the like. Such a controller may thus be regarded as a hardware component and the means for performing the various functions included therein may also be regarded as structures within the hardware component. Or even means for performing the functions may be conceived to be both a software module implementing the method and a structure within a hardware component.

The systems, apparatuses, modules or units described in the above embodiments may be specifically implemented by a computer chip or an entity, or implemented by a product with certain functions. One typical implementation device is a computer. In particular, the computer may be, for example, a personal computer, a laptop computer, a cellular telephone, a camera phone, a smartphone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

For convenience of description, the above devices are described as being divided into various units by function, respectively. Of course, the functionality of the various elements may be implemented in the same one or more software and/or hardware implementations in implementing one or more embodiments of the present description.

As will be appreciated by one skilled in the art, the present specification embodiments may be provided as a method, system, or computer program product. Accordingly, embodiments of the present description may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, embodiments of the present description may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and so forth) having computer-usable program code embodied therein.

The description has been described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the description. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

In a typical configuration, a computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include forms of volatile memory in a computer readable medium, Random Access Memory (RAM) and/or non-volatile memory, such as Read Only Memory (ROM) or flash memory (flash RAM). Memory is an example of a computer-readable medium.

Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

It should also be noted that the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising an … …" does not exclude the presence of other identical elements in the process, method, article, or apparatus that comprises the element.

This description may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.

All the embodiments in the present specification are described in a progressive manner, and the same and similar parts among the embodiments are referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, as for the system embodiment, since it is substantially similar to the method embodiment, the description is relatively simple, and reference may be made to the partial description of the method embodiment for relevant points.

The above description is only an example of the present disclosure, and is not intended to limit the present disclosure. Various modifications and changes may occur to those skilled in the art to which the present application pertains. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included in the scope of the claims of the present application.

Claims

1. A method of structured data classification, the method comprising:

2. The method of claim 1, wherein performing a sampling operation on one or more tables stored within a structured database to obtain a set of fields to which the tables correspond comprises:

the structured database comprises a distributed database adopting a block chain technology, a table in the structured database is obtained from a node of a block chain, the authorization state of the table is determined, and sampling operation is performed on the authorized table to obtain a field set corresponding to the table;

the table at least comprises a field of one type, and the field set comprises fields of preset line numbers corresponding to each type of field.

3. The method of claim 1, wherein the field information includes a value of a field and annotation information of the field, and the performing a scoring operation on the field information by using a pre-configured scoring rule or a scoring model to obtain a scoring result corresponding to the field comprises:

And taking the value of the field corresponding to each field and the comment information of the field as input, scoring the input by using a pre-configured scoring rule or a pre-configured scoring model to obtain scoring results corresponding to each field, and uploading the scoring results to nodes in the block chain.

4. The method of claim 1 or 3, wherein the preconfigured scoring rules or scoring models are configured according to categories of known fields; wherein, the first and the second end of the pipe are connected with each other,

5. The method of claim 4, further comprising:

when the field information is marked by adopting a marking rule, a regular expression is configured for the annotation information corresponding to each field type, and a rule set is generated according to the regular expression corresponding to the field and the regular expression of the annotation information corresponding to the field.

6. The method of claim 4, further comprising:

when the scoring model is used for scoring the field information, the values of the fields and the annotation information of the fields are used as parameters to be input into the scoring model, so that the scoring model can calculate a scoring function according to the parameters to obtain a scoring result.

7. The method of claim 4, wherein the scoring rules further comprise a list of keywords, and wherein the scoring operation is performed on the field information using a list of keywords preconfigured according to the category of the known field.

8. The method of claim 1, further comprising, before the training the predetermined deep learning model according to the scoring result corresponding to the field:

integrating the scoring results of multiple dimensions with the field set to obtain a data set, dividing the data set into a training set and a testing set, and training the preset deep learning model by using the training set; and each field in the field set corresponds to one scoring result.

9. The method of claim 1, wherein the predetermined deep learning model is a pre-constructed codec, and the codec is a custom model based on a deep learning framework; wherein the content of the first and second substances,

10. The method of claim 9, wherein the loss function is set to a mean square error between an input of the first concealment layer and an output of the third concealment layer during training of the codec.

11. An apparatus for structured data classification, the apparatus comprising:

12. A structured data sorting apparatus comprising:

at least one processor; and (c) a second step of,

13. A computer readable medium having computer readable instructions stored thereon which are executable by a processor to implement the structured data classification method of any one of claims 1 to 10.